The present invention relates to computing systems, and deals more particularly with performance management in such computing systems, whereby performance data capture is automatically and dynamically enabled.
Many software products provide the ability to capture data pertaining to system performance and throughput. There is a trade-off to gathering such data, however, since the act of doing so is a performance and throughput cost to the system. That is, CPU usage will increase due to gathering and capturing the data, and the elapsed time for applications executing in the system will increase, resulting in reduced throughput.